[gmx-users] Negative w problem

[Erik Lindahl]

Upgrade to 3.2(.1). I improved the CG minimizer significantly (even in 
single prec.) and added an even better L-BFGS method.

Or, if you just want to minimize a system to run dynamics steepest 
descent is more than enough.

Cheers,

Erik


On Apr 16, 2004, at 5:55 PM, Hua Wong wrote:

> During a CG minimisation, I obtain the error :
>
> Negative w:  -2.046612587870e+15
> z=  -2.548360507670e+07
> gpa=   1.347161368257e+07, gpb=   2.001264866334e+08
> a=   0.000000000000e+00, b=   2.602946975128e-08
> EpotA=  -1.441275875024e+04, EpotB=  -1.441068436024e+04
> Negative number for sqrt encountered (-2046612587870065.500000)
> Terminating minimization
>
> I read the mailing list about problems like this, and found something 
> about double precision. But I do use double precision gromacs.
>
> Any hint about the negative w ? It is the first time I encounter it.
>
> What are the possible problems?
>
> Thanks
> _______________________________________________
> gmx-users mailing list
> gmx-users at gromacs.org
> http://www.gromacs.org/mailman/listinfo/gmx-users
> Please don't post (un)subscribe requests to the list. Use the
> www interface or send it to gmx-users-request at gromacs.org.